Oxidations on DSA and chirally modified DSA and tin dioxide

Dec 1, 1976 - Bruce E. Firth, Larry L. Miller. J. Am. Chem. Soc. , 1976, 98 ... William R. Heineman and Peter T. Kissinger. Analytical Chemistry 1978 ...
0 downloads 0 Views 277KB Size
8272 (5) The spectra were obtained on the DuPont instrument at the University of North (6)

(7) (8)

(9)

Carolina. Results on silylated carbon electrodes are reported independently, C. M. Elliott and R. W. Murray, Anal. Chem., in press. The basal surface is not a prefect planar sheet. It is, therefore, possible to bind at dislocations as well as adsorb. This can account for the nitrogen signal. B. Firth and L. L. Miller. J. Am. Chem. Soc., following paper in this issue. Optical yields are lower using the edge surface than the spectroscopic rods. The reason(s) for this is under investigation, but most likely involves differences in the surfaces. Address correspondence to this author at the Department of Chemistry, University of Minnesota, Minneapolis, Minn. 55455.

Bruce E. Firth, Larry L. Miller*9 Michiharu Mitani, Thomas Rogers Department of Chemistry, Colorado State Unicersity Fort Collins. Colorado 80523

Table I. Optical Purity of Sulfoxidesa Electrode

Reactant

Modified DSA

la

Modified SnOz

la lb la

Potential

Enantiomeric excess

(V)

-2 (%)

1.4 1.1 1.1 I .8

0.9c 0.8 0.3 1.4

(’Oxidations performed as described in text. Measured vs. AgIAgNO3. Average of four runs. Deviation of a589 f0.1’. Ocobsd

-0.045

(C

3, CHCI?).

was covered with Dow Corning Silastic 732 RTV after functionalization. There are to our knowledge no reports of asymmetric, anodic reactions. The anodic process chosen for study was the conversion of an aryl methyl sulfide to the sulfoxide. This reaction is relatively simple and proceeds in high chemical yield.9 The sulfoxide molecules are chiral and the production of optically enriched product constitutes the test of successful modification. Oxidations on DSA and Chirally Modified DSA and The oxidation was carried out in a three-compartment cell using Ag(O.1 M AgNO3 in CH3CN as a reference electrode. SnOr Electrodes The electrolyte consisted of 15 g of tetraethylammonium fluSir: oroborate in 140 ml of acetonitrile and 10 ml of water.9 In a typical experiment on DSA the potential was held a t 1.4 V, The chemical modification of electrode surfaces has recently come under investigation in our laboratories. Our goal is to background current was 20 mA, initial current from 500 mg produce interfacial regions which can be used to perform of l a was 150 mA and after passage of 2 faraday/mol the final specific and unique reactions. Initial approaches have concurrent was 30 mA. On S n 0 2 a t 1.8 V a typical background centrated on synthesizing chiral electrode surfaces by binding was 1 mA, initial current from 500 mg of l a was 28 mA and electroinactive, chiral compounds to conductors. Such chiral final current 2 mA. The products were isolated by evaporation electrodes are then used to perform asymmetric electrode reof the anolyte, followed by addition of 100 ml of water and actions. This approach was selected because asymmetric extraction with two 100-ml portions of ether. Rotary evaposynthesis requires a chiral reagent. It, therefore, provides a ration after drying the ether over anhydrous MgS04 gave the sensitive probe for a successful modification which remains crude sulfoxide (-350 mg, current yield 65%) which was puintact and “active” during use. The first reported example is rified by chromatography on silica gel using CHC13 as elutent or by preparative T L C using a 1: 1 mixture of chloroform and the preparation of “(S)-C,IPheM” involving binding (S)ethyl acetate as developer. All oxidations were run a t room (-)-methyl phenylalanate to a carbon electrode via the surface temperature under air (no change was observed if the reaction oxides.’.2 This material was used to produce optically active was run under argon or nitrogen) with the solution stirred with alcohols by the reduction of ketones.[ a magnetic stirrer. In the present study we have prepared chiral surfaces from The enantiomeric excesses shown in the table are of a low DSA3 and SnOz electrodes and used them for preparative, magnitude,lO but they are quite reproducible and further puasymmetric oxidations. Evidence is available which indicates that meta oxides can be chemically modified by ~ i l y l a t i o n , ~ . ~rification by chromatography gave sulfoxide with an unchanged rotation. Crystallization gave material with higher but these materials have not been used for preparative elecrotations as is common for such compounds.” Rotations were trolysis. It was hoped that “active” and stable surfaces could taken a t four wavelengths. Partially resolved material was be prepared which would be useful for anodic synthesis. compared to the electrolysis product and direct proportionality To our knowledge DSA electrodes have not been previously between all readings was obtained. The optical purity of the used for preparative organic electrochemistry and we have, product was established by N M R using tris[3-(trifluorotherefore, investigated the use of this material without surface methylhydroxymethylene)-d-camphorato]europium(III). The modification. It is found that the onset of background in methyl signal of a partially resolved sample split into a doublet C H I C N containing 4% HzO is about 1.8 V vs. AglAgN03 in and allowed the specific rotation of pure 2a to be assigned as CH3CN. The surface is quite stable in this solvent a t 1.5 V, and (r20589 167O. The rotation of optically pure 2b is a20589 146O.I2 very high current densities are supported. Several reactions, Although the selectivity is not useful for asymmetric synthesis, e.g., cyclization of laudanosineh and conversion of l-anisylit is comparable to the results obtained in similar conversions ethanol to p-methoxyacetophenone, proceed with yields using chiral chemical reagents.” W e have made a specific equivalent to those obtained using platinum. Therefore, these comparison using l a with (+)-percamphoric acidt3 which electrodes seem to be a promising addition to the limited list produced 2a with enantiomeric excess 1.4. of useful anodes. To test the possibility that an adsorbed chiral reagent would Modification of antimony doped SnO2 on a glass backing be effective, a comparison experiment was conducted by oxiand DSA on a titanium backing was achieved by the following dizing l a under the same conditions as above except that an procedure. The electrode (general dimensions 30 X 80 mm) unmodified DSA electrode was used and (+)-camphoric acid was placed in 100 ml of dry benzene (distilled from C a H ) (equimolar with l a ) was added to the anolyte. An electrode containing 3 ml of y-aminopropyltriethoxysilane.The elecwhich had only been silanized was used in separate experiments trode was transferred after 30 min’ to a dry benzene solution with and without camphoric acid. In each case 2a was isolated of (-)-camphoric anhydride (2 g in 100 ml). After 24 h the electrode was removed, washed with acetone, and was ready but was optically inactive. An attempt was also made to directly bind (+)-camphoric anhydride to DSA by soaking the for use.8 I n the case of the DSA electrodes the titanium backing John Lennox, Royce W. Murray Department of Chemistry, Uniuersity of North Carolina Chapel Hill, North Carolina 2751 4 Received July 7 , 1976

+

Journal of the American Chemical Societj! / 98:25

/

December 8, 1976

8273 electrode in a benzene solution of the anhydride for 24 h. The resulting material was used under the usual conditions to convert l a to 2a. The product was again optically inactive. 0

1

2

a,X=NO, b, X = CH,

Other experiments showed that the modified DSA electrode gave identical optical yields in three consecutive runs and that there was little dependence of optical yield on the potential used. There are, however, differences in the voltammograms of l a on modified and unmodified surfaces. These will be reported on and analyzed in a full publication.

Acknowledgment. Support by the National Science Foundation is gratefully acknowledged. Materials were supplied by Professor T. Kuwana, Dr. R. Neale, and Electrode Corporation. Stimulating discussions with Professor Kuwana and Dr. Neale are also acknowledged. References and Notes (1)B. Watkins, J. Behling, E. Kariv, and L. L. Miller, J. Am. Chem. Soc., 97, 3549 (1975). (2)The accompanying communication reports on the use of C,IPheM in asymmetric sulfoxide synthesis. A variety of amino acids have been attached to carbon. Unpublished work of M. Mitani from this laboratory.

(3)These DSA electrodes contain tantalum and iridium oxides. They were supplied by Electrode Corporation, Chardon,Ohio. (4) N. R. Armstrong, A. W. C. Lin, M. Fujihara, and T. Kuwana, Anal. Chem., 48, 741 (1976). (5) P. R. Moses, L. Wier, and R . W. Murray, Anal. Chem., 48, 1882 (1975). (6)L. L. Miller, F. R . Stermitz, and J. R . Falck, J. Am. Chem. Soc., 93, 5941

(1971). (7)This is essentially the procedure used by Murray and co-workers for SnOz. It was originally suggested to us by R . Neale, Union Carbide Corporation. In the present work it was found that a 24-h reaction period gave an electrode which passed very little current under the usual conditions. This is presumably due to polymer formation. (8) Although there is some logic in this selection of reagents, it is deficient in two aspects. The use of camphoric anhydride means that two diastereomeric products could result. Second, it has been very recently suggested by R. W. Murray, based on ESCA measurements, that only a fraction of the amine groups on "y-aminopropylated-SnOz" are available for reaction. (9)V. D. Parker in "Organic Electrochemistry", M. Baizer, Ed., Marcel Dekker, New York, N.Y.. 1973,p 551. (10) A control experiment in which l a was oxidized in the presence of partially resolved 2a established that the latter was partially racemized during the oxidation. This apparently takes place at the surface of the polarized electrode since there is no racemization in solution without electrolysis. ( 1 1 ) J. D. Morrison and H. S. Mosher in "Asymmetric Organic Reactions", Prentice Hall, Englewood Cliffs, N.J., 1971,p 335. (12)K. Mislow, M. M. Green, P. Law, J. T. Melillo, T. Simmons, and A. L. Ternay, Jr.. J. Am. Chem. SOC.,87, 1958 (1965). (13)J. F. Collins and M. A. McKervey, J. Org. Chem., 34, 4172 (1969). (14)Address correspondence to this author at the Department of Chemistry, University of Minnesota, Minneapolis, Minn. 55455.

Bruce E. Firth, Larry L. Miller* l4 Department of Chemistry, Colorado State liniversity Fort Collins, Colorado 80523 Received July 7 , 1976

Carbanion Chemistry. A Room Temperature Retro-Diels-Alder Reaction Sir; The retro-Diels-Alder reaction commonly requires temperatures of 150 "C or more for convenient reaction.'$2The extensive chemistry of norbornene and norbornadiene provides ample demonstration of the resistance to t h e r m ~ l y s i s .In ~ striking contrast we now report that the 7-phenylnorbornenyl anion I undergoes the retro-Diels-Alder reaction within 30 min

7.5

7.0

6.5

55

6.0

PPM

Figure 1. The 100-MHz ' H NMR spectrum of phenylcyclopentadienide ion in DME-die. Chemical shirts are shown in d units from internal Me&.

a t room temperature. When a 1:3 mixture of syn- and anti7-phenyl-7-metho~ynorbornene~ was stirred at room temperature with excess sodium-potassium alloy in 1,2-dimethoxyethane in an evacuated r e a ~ t o r gas , ~ evolution was observed and a deep green6a solution formed over a period of 30 min. The gas was collected and mass spectrometric analysis identified it as ethylene contaminated with solvent, 1,2-dimethoxyethane. The yield of ethylene was estimated to be 6W0 from mass spectrometric analysis. An N M R sample of the filtered solution in DME-dlo showed a clean spectrum of the phenylcyclopentadienide ion, I1 (Figure l ) , as the only other species present. A I3CN M R spectrum6bconfirmed this conclusion. It is noteworthy that neither of the starting materials nor any other products were detected within the 5% limits set by sensitivity criteria. The structure of ion I1 was deduced from the characteristic phenyl anion absorptions and the AA'BB' pattern of a monosubstituted cyclopentadiene in the proton spectrum. Proton decoupling confirmed the coupling interactions. When the anion solution was poured into ethanol, the color immediately discharged, but the transparent mixture darkened and turned black within 5 min. No low molecular weight product was obtained from a subsequent ether extract. This observation is consistent with the known chemistry of phenylcy~lopentadienes.~

g+ P

-

C2H4

I1

In an analogous pair of reactions,syn- and anti-9-phenyl,9methoxybenzonorbornene were separately cleaved a t -40 O C by Cs/K/Na alloy.* In 2 h a t 0 O C gas evolution was complete yielding a clean solution of the 2-phenylindenide anion as the sole product. The proton spectrum showed a singlet 6.20 ( H l ) , an AA'BB' pattern 6.30, 7.00 (H4, H5, respectively), and the characteristic phenyl anion pattern at 6.78 (Hp), 7.15 ( H m overlapping H5), and 7.62 (Ho). Methanol quench yielded 2-phenylindene. These observations now provide a rational explanation for several literature reports. Treatment of norbornadiene with amyl sodium gave a quantitative yield of the cyclopentadienyl anion and acetylene in 5 h.l' When 7-tert-butoxynorbornadiene was cleaved with sodium/potassium alloy, only a 20% Communications to the Editor